LHC@home

Challenge issued on BOINCstats/BAM starting on 26/02/2015 00:00 UTC issued by The Scottish Boinc Team. One week of crunching on LHC@Home. All teams are welcome to sign up and join us on this challenge.
Hope to see you there
Alez [ TSBT's Pirate ]

http://boincstats.com/en/stats/challenge/team/chat/649

Alez,

I am new to LHC at home. I like the idea of a bit of competition.
Been working on other project but this is the 1st time I have seen this function within Boinc.

Great idea. Servers at the Ready so to speak! ;-)

Steven

---

• Cloud Storage Reviews
  
• Fullsus.com Mountain Bike Site
  
• UHDTV Comparisons and Reviews
  

I hope they'll have enough work available for your competition.
Work flow isn't guaranteed here and tends to come in batches so I hope your team-mates won't be disheartened if some get work and others don't

All we can hope is that there is work available. Fullsus, enjoy the challenge and pop by our forum for the team competition chat.

I have asked physicists to make an effort to
schedule work for this challenge. Eric.

---

We appreciate your help Eric. Stop by TSBT forum sometime http://www.thescottishboincteam.co.uk

TSBT forum is open to everyone :)

Having fun with the Challenge! Hoping when I get home to night I'm able to pull some more WU's, not too afraid looking at the server status but was concerned after seeing the tower only had eight more to chew on before OoW.

Had a quick question for any that may know. Been looking through the WU's my tower has completed and see that my wingman has used different instruction sets on their WU during validation. Was just curious if this is just what is expected with different processors? Will try and find some WU's for examples here in a bit and update my post.


Examples:

Some of these are Intel/AMD and some are AMD/AMD but different instruction set used for validation.
http://lhcathomeclassic.cern.ch/sixtrack/workunit.php?wuid=28353511
http://lhcathomeclassic.cern.ch/sixtrack/workunit.php?wuid=28348321
http://lhcathomeclassic.cern.ch/sixtrack/workunit.php?wuid=28348322

Doh! Answered my own question...
http://lhcathomeclassic.cern.ch/sixtrack/forum_thread.php?id=3888#26916

'PNI' stands for Prescott New Instructions, and was Intel's trade name for what became SSE3 (http://en.wikipedia.org/wiki/SSE3). So the instruction sets are the same, and the same application has indeed been deployed twice so that BOINC's feature detection works correctly on both Intel and AMD CPUs.

I don't know the thinking on why a separate 64-bit deployment was necessary too.

The units are flowing nicely. Thanks Eric.

Next round of Q's:

There are currently 5 days left of the challenge. Is there the ability to produce more Wu's for those that are running out?

Server updates? Took a few looks over the past hour and saw 1 WU ready to send and 2 waiting validation....Personal count of pending are 11 for my own account, was wondering if this is tracking something other then account pending?

Thanks for the help.

I decided to stop by and see how things are going and see we added some COLOR to the LHC site.

---

Volunteer Mad Scientist For Life

I was looking at who was crunching the most during this challenge and realized the top LHC cruncher wasn't even part of the challenge, or a team. So who is/are the Agile Boincers? Is that the Worldwide LHC Computing Grid?

Competition now over. Thanks to Eric and all the physicists for keeping the work flowing over the course of the event. I hope our efforts prove useful to you.

Well that has been worthwhile. We got a lot of
work done and discovered a couple of problems
with our setup. Very tricky but it is my priority
to solve them. One difficult issue seems to be
different results from Windows and Linux (perhaps),
but only about 6 out of 20,000 that I know of.
It is difficult because I don't have all combinations
of Linux Windows Intel Athlon available for testing.
I shall have to search the validator logs (huge) and try and
figure out what is wrong. The other is that we are getting
back empty result files, not a problem on our side
and I think credit is given anyway.

This exercise has also made me think about switching to
gfortran instead of Intel ifort; this would seem to
be appropriate for a volunteer project. I need to check the
performance numbers and also find the appropriate gfortran flags
for code generation. (Actually I am getting identical
results from 4 different Fortran compilers, at different degrees
of optimisation, on Windows Linux and Mac, but have a
problem with generating the Mac executable....:-( .) Eric.

---